Menshevising Idealism: or

why the Soviet Union didn't

develop the first computers
Greg Michaelson
Oxford Communist Corresponding Society
3rd December 2020
Overview
• let’s start with some context
• 1946 - first modern digital computers announced
• ENIAC – USA – University of Pennsylvania – in use
• ACE – UK – National Physical Laboratory - design
• 1948 - many early US & UK computers running
• 1951 - commercial computers
Overview
• 1951 - first comparable Soviet computer running
• MESM – Sergey Lebedev
• why...?
• USSR had theoretical/technical capability
• control theory
• analogue computing
• sequence control systems
• computers key to economic planning & automation
Overview
• digital computers not secret
• long history of mechanical computation
• mechanical calculators
• 1642 - Pascal - France - adder/Pascaline
• 1694 - Leibniz - Germany – multiplier
• 1822 - Babbage – UK - difference engine – generate tables
• analogue computers
• WWI - Great Powers - naval fire control
• 1928 – Vannever Bush – USA - differential analysers
• 1936 – Lukyanov - USSR- water integrator
Overview
• proto digital computers
• 1936 – Zuse - Germany - Z1 - patents 1937
• 1939 – Atanasoff – USA - ABC
• 1944 - USA - ASCC/Harvard Mark 1
• secret proto digital computers
• 1943 - Colossus – UK
• used to crack German Lorenz cipher
• not public until 1980s
• techniques relevant to analysing streamed data
Overview
• 1945 - von Neumann Report on the EDVAC
• described stored program design – key to digital computers
• widely distributed
• all subsequent general purpose digital computers stored program
• why didn’t USSR pick up on this?
• anecdotally, Lebedev read about digital computers in 1948 in foreign popular
magazines
• why didn’t USSR prioritise acquiring US/UK designs?
• c.f. atomic weapons – 1940s
What’s special about computers?
• computers are machines that execute instructions to transform data
• program – sequence of instructions
• two components
• Central Processing Unit
• carries out instructions
• memory – holds data
• include peripherals (e.g. mouse, keyboard, screen) & network
• instructions to
• move data from memory to CPU
• change data in CPU – arithmetic & comparison
• move data from CPU to memory
What’s special about computers?
• usually instructions are carried out in a linear sequence
• also instructions to change the sequence depending on properties of
data
• choose between optional sequences
• repeat a sequence until some property is satisfied
• proto digital computers instruction sequences fixed
• had to be physically changed for each program
• ENIAC – Electronic Numerical Integrator and Computer
• functional components wired together
• Colossus
• instructions set on switches and jack panels
What’s special about computers?
• modern digital computers store instructions in memory
• as well as data
• easy to change what computer does
• computer can modify its own programs
• stored program attributed to John von Neumann
• modified ENIAC – 1946/47
• von Neumann got idea from Alan Turing - 1936/7
What’s special about computers?
• Turing was working on mathematical logic
• theoretical foundations of mathematics
• concerned with computability
• what are limits to mechanical rule following?
• NB well before digital computers
• let’s now survey mathematical logic before considering its
implications for philosophy
Logic and computability
• propositional calculus
• Aristotle - 384-322 BC - Greece
• values: true & false
• variables over values: x y z...
• operators
• conjunction – logical and – x AND y - true only if both x and y true
• disjunction – logical or – x OR y - true if either x or y or both true
• negation – logical not – NOT x - true if x false
• implication – x implies y or if x then y
Logic and computability
• expressions – built up using truth values, variables and operators
• theorem – expression that is true for all values of variables
• axiom – basic true expression
• rule of inference
• syllogism
• connects premises to conclusion
• implication
Logic and computability
• proof – establish something is a theorem
• start with axioms as premises
• apply rules of inference
• until expression is conclusion
• deductive logic
• NB
• pure calculus – doesn’t matter what variables represent
• form of logic Hegel, Marx & Engels familiar with
Logic and computability
• predicate calculus
• Augustus De Morgan - 1806-1871 - UK
• extends propositional calculus
• predicates – over truth variables
• properties that may be true or false – P Q R...
• quantifiers
• universal – for all
• existential – there exists
• more axioms and rules of inferences
• NB still a pure calculus – doesn’t matter what predicates represent
Logic and computability
• set theory
• Cantor – 1845-1918 – Germany
• set is a collection of things
• symbols or other sets
• operators
• member – some thing a member of a set
• union – join two sets – no duplicates
• intersection – common elements of two sets
• difference – between two sets
Logic and computability
• apply propositional & predicate calculus to set theory
• predicates are properties of sets
• quantify over sets
• e.g. model arithmetic
• but
• countable infinite sets
• Engels unhappy
• mocks Cantor’s ideas in Anti-Duhring
• paradoxes
• Russell’s
• set of all sets that aren’t members of themselves
• is this set a member of itself?
Logic and computability
• can logic capture all of mathematics?
• 1910-27 - Russell & Whitehead – Principia Mathematica – UK -
Cambridge
• heroic failure
• David Hilbert – 1862-1943 - Germany
• Hilbert’s Programme
• can we prove for mathematics
• consistency - no provable contradictions
• completeness – no unprovable theorems
• decidability – systematic way of establishing whether or not an expression is a
theorem
Logic and computability
• Kurt Gödel – 1906-1978 – Austria/USA
• 1931 - mathematics can’t be both consistent and complete
• Alan Turing – 1912-1954 – UK - Cambridge
• 1936 - mathematics can’t be decidable
• proofs by contradiction
Logic and computability
• Turing’s distinctive approach
• most logical models of maths require an implicit human to do the proofs
• Turing invented a class of machines based on how humans do mathematics
• tape
• unbounded but finite sequence of symbols
• rules
• in current state
• depending on current symbol
• change symbol
• move tape one symbol left or right
• change state
Logic and computability
• NB machine - no implicit human
• Turing showed
• any mathematical expression can be represented on the tape
• Turing machines can be represented on the tape
• can make a Universal Turing machine to execute Turing machines
➔ Von Neumann’s stored program
Logic and computability
• what is the ontological status of logic?
• logicism/logical atomism e.g. Russell
• mathematics is reducible to logic
• truth is out there
• Platonic realism i.e. idealism
• formalism e.g. Hilbert
• mathematics is applying rules to symbols – language game
• no necessary meanings
Logic and computability
• intuitionism e.g. Brouwer - Netherlands
• only accept mathematical entities that can be constructed from first
principles
• rejects actual infinities
• rejects proof by contradiction
• led to constructivism e.g. Markov (Jnr) – USSR
• let’s now look at how this all played with Marxist philosophy
Menshevising idealism
• mathematical logic seen as contrary to dialectical materialism
• mechanical materialism
• formalist
• roots in Engels on logic and mathematics
• dialectical materialism == logic + dialectics
• pure dialectics == Hegelian idealism
• pure logic == mechanical materialism
Menshevising idealism
• 1878 - Anti-Dühring
• philosophy of nature is empirical, abstracted from reality
• dialectics goes beyond “narrow horizon” of formal logic
• 1883 - Dialectics of Nature
• formal logic enumerates “forms of motion of thoughts”
• dialectics derives higher forms from lower
• formal logic is not nonsense
• but pure maths is abstractions which becomes nonsense at extremes
• 1886 - Ludwig Feuerbach and the End of Classical German Philosophy
• all that’s left of philosophy “expelled from nature and history” is theory of
laws of thought –logic and dialectics
Menshevising idealism
• struggle over philosophy of science in USSR after October Revolution
• 1920s - mechanist tendency gained ascendency
• down played dialectics
• reductionist
• Abram Deborin
• dialectical materialism is a synthesis of Hegelian dialectic and Feurbach
materialism
• anti reductionist
Menshevising idealism
• 1929 – mechanists discredited as revisionist
• purge of Academy of Sciences
• 1930 – New Turn
• V. P. Mitulin accused Deborin of formalism in approach to dialectics
• not relevant to practical needs
• Stalin coined “menshevising idealism” for Deborin
• by 1931, mechanism & MI conflated with political orientation
• mechanism == right deviation – Bukharin
• MI== left deviation - Trotsky
Menshevising Idealism
• made career progression difficult...
• led to Soviet focus on applicable mathematics
• subsequent Soviet mathematical logic based on
intuitionism/constructivism
Menshevising Idealism
• international implications
• e.g. UK
• CPGB had leading scientists as members
• 1931 - Second International Congress of History of Science &
Technology
• London
• paper by Hessen on Soviet line
Menshevising Idealism
• David Guest - 1911-1938
• mathematician – Cambridge & Southampton
• killed in Spain
• supported Hessen at Congress
• 1934 - Machian Tendency in Modern British Philosophy
• orthodox Soviet position
• attacked logicists and formalists
• promoted intuitionism as revolutionary
Menshevising Idealism
• 1934 – collection - Aspects of Dialectical Materialism
• two papers by communist scientists
• Hyman Levy - 1889-1975
• mathematician- Imperial College
• friend of David Guest
• laws of dialectics add nothing to science
• subsequently criticised as mechanist
• J. D. Bernal - 1901-1971
• X-Ray crystallographer - Cambridge & Birkbeck
• dialectic is not applicable to experimental science
Menshevising Idealism
• J. B. S. Haldane – 1892-1964
• geneticist/physiologist - UCL
• 1938 - Marxist Philosophy and the Sciences
• orthodox Soviet position
• reiterated Engels’ disquiet with infinite sets
• Marxists sceptical of “more ambitious logical theories”
• logical systems only work in highly abstract contexts
• knew about calculating machines – differential analysers
• said may herald a “new epoch” in mathematics
Menshevising Idealism
• Maurice Cornforth – 1909-1980
• logician/philosopher
• 1946 - Science Versus Idealism
• orthodox Soviet critique of pure empiricism and modern logic
• tribute to David Guest
• helped by Haldane
Menshevising Idealism
• these British communists all knew each other...
• ... and had
• close contact with USSR
• strong links to currents in logic and mathematics
• especially Cambridge
• none seem to have noticed Turing’s papers
Conclusions
• lost opportunity for USSR
• compounded by 1940s critique of cybernetics
• e.g. 1950 Boris Agapov - editor Literary Gazette
• capitalists want to replace class conscious workers & soldiers with compliant
robots
• computers no use for planning
• but Soviet military & scientists wanted computers
• two secret design programs
• Academy of Sciences - MESM
• Ministry of Machine-Building and Instrument Construction - Strela
Conclusions
• easier after 1953 “thaw”
• mid 1960s - USSR could build world class computers e.g. BESM 6
• west worried by progress
• CIA report suggested USSR might overtake
• too many competing projects
• from different organisations – ministries, academia, military
• no standardisation of hardware and software
• couldn’t ramp up to meet demand
• late 1960s - own designs abandoned for IBM 360 clones
Conclusions
• Turing not ideological/political
• irony that approach was materialist in Engels’ terms
• from observation to theory to practice
• Turing’s results constrain computation
• constrains reach of logic
• Turing also argued that thought is computational
• so we might conclude that dialectics as a component of thought is
constrained
• worth exploring implications for dialectical materialism
• relationships between computation, logic and dialectics
Further reading
• Helena Sheehan, Marxism and the Philosophy of Science: A Critical
History, Humanities Press, 1985, 1993, Verso Books, 2017.
• Paul Cockshott, Louis Mackenzie and Greg Michaelson, Computation
and its Limits, Oxford University Press, 2012